CN112085727A - Intelligent identification method for scale structure on surface of hot rolled steel - Google Patents
Intelligent identification method for scale structure on surface of hot rolled steel Download PDFInfo
- Publication number
- CN112085727A CN112085727A CN202010978690.3A CN202010978690A CN112085727A CN 112085727 A CN112085727 A CN 112085727A CN 202010978690 A CN202010978690 A CN 202010978690A CN 112085727 A CN112085727 A CN 112085727A
- Authority
- CN
- China
- Prior art keywords
- rolled steel
- scale
- hot rolled
- hot
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides an intelligent identification method for a scale structure on the surface of a hot rolled steel product, and relates to the technical field of steel rolling. According to the method, a sample for detecting the oxide scale is prepared, metallographic detection equipment is utilized to obtain the photo data of the oxide scale on the surface of the hot-rolled steel, and an oxide scale image sample set is established after pretreatment. And (3) making a semantic label, training the neural network model by combining with a neural network model for establishing the semantic analysis of the scale structure image, and finally realizing that a user inputs a scale tissue picture and automatically acquires the proportion, the distribution area and the morphological classification description of each tissue in the picture.
Description
Technical Field
The invention relates to the technical field of steel rolling, in particular to an intelligent identification method for an iron scale structure on the surface of a hot rolled steel product.
Background
With the rapid development of the manufacturing industry in China, the requirement on the surface state of steel is rapidly improved, and steel enterprises are forced to regard the surface state of products as a key index for measuring the quality of the products, so that the high surface quality control technology also becomes a core competitiveness of the enterprises. More than 98% of the hot-rolled carbon steel is Fe element, and the content of other alloy elements is very low, so that the oxide of the final product is Fe oxide (FeO, Fe)3O4Etc.) and Fe/Fe generated in cooling3O4Eutectoid structure is the main, and different effects of hot rolling production can be realized through different oxide compositions. As the key content in the field of surface quality control, accurate differentiation of the scale structure of hot-rolled carbon steel becomes an important prerequisite for improving surface quality. The image scheme needs to be calibrated from the traditional mode by manual work to each part of tissue, and then statistics and summation are carried out, so that the mode is complex in operation, low in efficiency, labor-wasting and difficult to adapt to the requirement of large-range use. With the advance of technology and the exponential improvement of computer hardware performance, image recognition methods based on neural networks are becoming mainstream. This technology has also found some applications in various fields. If the method is applied to industry and obtains good effect, the method can be used for bricking and tiling the digitally transformed roads in China.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an intelligent identification method for an iron scale structure on the surface of a hot rolled steel product, which aims to solve the problems that the iron scale tissue identification efficiency is low and accurate statistics is difficult.
The technical scheme adopted by the invention is as follows:
an intelligent identification method for a scale structure on the surface of a hot rolled steel product comprises the following steps:
step 1, manufacturing a detection sample of hot rolled steel oxide scale, shooting by using metallographic detection equipment to obtain a hot rolled steel oxide scale picture, and establishing a hot rolled steel oxide scale image sample set after pretreatment;
the manufactured detection sample is subjected to hot inlaying, mechanical polishing and chemical corrosion to remove stress on the oxide scale of the hot rolled steel;
the metallographic detection device is in a backscattering mode of a scanning electron microscope or an electronic probe;
the pretreatment is to adjust the matrix part in the hot rolled steel iron oxide scale picture to be standard white of an RGB color difference space;
step 2, making a structure semantic label graph of each hot-rolled steel iron oxide photo in the marked hot-rolled steel iron oxide image sample set;
step 2.1, Fe of hot rolled steel iron scale picture3O4FeO and eutectoid tissues are subjected to regional division, so that the iron scale phase tissues in the same region are ensured to be the same;
step 2.2, dyeing each area, wherein the embedding material part, the matrix part and the picture shooting information part in the picture are used as background processing and are marked to be black, and the dyed positions are independently output to obtain a tissue semantic label picture;
the dyeing treatment is that the same tissues in the hot rolled steel iron oxide scale pictures on the tissue semantic label pictures of all the sample sets are dyed into the same color;
step 2.3, carrying out image size normalization processing on the structural photo of the iron scale and the semantic tag graph of the corresponding iron scale tissue;
step 3, constructing a neural network model for semantic analysis of hot rolled steel oxide scale images, training the neural network model by using the hot rolled steel oxide scale image sample set in the step 1 and the organization semantic label graph in the step 2, and setting the training times for N times to obtain model parameters;
step 4, user directionInputting the trained neural network model into a hot rolled steel iron scale picture to be detected and shot by metallographic detection equipment, and automatically acquiring Fe in the input hot rolled steel iron scale picture by using the neural network model3O4The proportions, distribution regions, and morphological classifications of FeO and eutectoid structures are described.
Step 4.1: reading a jpg image file of the scale of the hot rolled steel to be detected, which is shot by metallographic detection equipment, in a specified folder, and converting the image file into a form of RGB data;
step 4.2: recognizing a neural network model by using the iron oxide scale structure trained in the step 3, outputting each pixel point of the iron oxide scale structure in the step 4.1 as a feature vector by the last layer of network through a normalized index function softmax, then matching a label corresponding to the component with the maximum output vector value, and outputting label file information to obtain a label graph of the image to be recognized, namely the distribution condition of each part of the structure;
the scale structure recognition neural network model is set to be 8 layers, wherein the first 5 layers are convolution layers and used for extracting image features, and the last three layers are anti-convolution layers and used for recovering the image size and carrying out logic inference; establishing a pooling layer in the convolution process, and adopting a maximum pooling method of 3-by-3 pooling windows, wherein the step length of the pooling layer is 2, and the pooling layer is used for fusing features and reducing the dimension of the image;
the iron scale structure recognition neural network model adopts ReLU as an activation function, and the expression is as follows:
F(x)＝max(0,x)
wherein x represents a pixel in an RGB channel in the convolution layer, and represents an output value of a neuron in an iron oxide scale structure recognition neural network; inputting the image data into the established network model for training, continuously optimizing the structural parameters of the network model in the training process, and finally storing the training parameters as a binary file.
Adopt the produced beneficial effect of above-mentioned technical scheme to lie in:
the invention provides an intelligent identification method for an iron scale structure on the surface of a hot-rolled steel product, which is used for identifying and counting the iron scale structure by utilizing a neural network, realizing accurate distinguishing of the iron scale structure of hot-rolled carbon steel, optimizing a hot-rolling process on a rolling production site and providing assistance for improving the surface quality of a product.
Drawings
FIG. 1 is a flowchart of an intelligent iron scale structure identification method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a scale sample to be detected according to an embodiment of the present invention;
FIG. 3 is an image of an iron scale tissue photographed and color-adjusted in an electron microscope backscattering mode according to an embodiment of the present invention;
FIG. 4 is a label image after tissue calibration and staining according to an original scale image in an embodiment of the present invention;
FIG. 5 is an image of a scale texture to be identified according to an embodiment of the present invention;
fig. 6 shows a picture recognition result according to an embodiment of the invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings.
An intelligent identification method for a scale structure on the surface of a hot rolled steel product is shown in figure 1 and comprises the following steps:
step 1, manufacturing a detection sample of hot rolled steel oxide scale, shooting by using metallographic detection equipment to obtain a hot rolled steel oxide scale picture, and establishing a hot rolled steel oxide scale image sample set after pretreatment;
the manufactured detection sample is subjected to hot inlaying, mechanical polishing and chemical corrosion, the iron scale of the hot rolled steel is destressed, the iron scale structure of the surface layer of the steel is clear, and the schematic diagram of the manufactured sample is shown in FIG. 2;
the metallographic detection device is in a backscattering mode of a scanning electron microscope or an electronic probe;
the pretreatment is to adjust the matrix part in the hot rolled steel iron oxide scale picture to be standard white of an RGB color difference space;
in the embodiment, iron scales are distributed on the upper side and the lower side of a sample matrix in the center of the circular mosaic material block; collecting the tissue image data of the iron scale by using a scanning electron microscope or an electronic probe, and taking an iron scale tissue image in an electron microscope back scattering mode as shown in fig. 3; and then establishing an iron scale tissue image data set by using the obtained image. The data set comprises training set image data and testing set image data; and performing secondary sorting on large-scale training data, screening 1000 pieces of training set image data in the embodiment, and selecting one fifth of the training set image data as verification set image data.
Step 2, making a tissue semantic label graph of each hot rolled steel iron oxide photo in a marked hot rolled steel iron oxide image sample set, marking the obtained tissue image data of the hot rolled steel iron oxide in a training set and a verification set, and marking the tissue image data of the hot rolled steel iron oxide in the verification set, wherein FIG. 4 is an image obtained by marking the tissue according to an original iron oxide image and carrying out calibration and dyeing;
step 2.1, Fe of hot rolled steel iron scale picture3O4FeO and eutectoid tissues are subjected to regional division, so that the iron scale phase tissues in the same region are ensured to be the same;
step 2.2, dyeing each area, wherein the embedding material part, the matrix part and the picture shooting information part in the picture are used as background processing and are marked to be black, and the dyed positions are independently output to obtain a tissue semantic label picture;
the dyeing treatment is that the same tissues in the hot rolled steel iron oxide scale pictures on the tissue semantic label pictures of all the sample sets are dyed into the same color;
step 2.3, carrying out image size normalization processing on the structural photo of the iron scale and the semantic tag graph of the corresponding iron scale tissue;
step 3, constructing a neural network model for semantic analysis of the hot rolled steel oxide scale image, training the neural network model by using the hot rolled steel oxide scale image sample set in the step 1 and the organization semantic label graph in the step 2, and setting the training times for N times to obtain high-precision model parameters;
building an FCN network model under a computer platform, and training by using an image data set to obtain training parameters of the model;
the experimental platform used by the implementation model is as follows:
hardware environment: the system takes a desktop computer which is loaded with an Intel (R) core (TM) i5-75003.40GHz processor and an internal memory 8G as a basic operating platform, and is configured with an NVIDIAGTX1070 video card with a video memory of 8G as acceleration equipment of the CUDA.
Software environment: the operating system used by the experimental platform was Ubuntu. The installed support libraries mainly comprise Python3.6, OpenCV-Python and Pythrch-1.
Through the size normalization process, a backhaul network of an FCN network model is then established under a deep learning framework, and the network model has 50 convolutional layers, which are denoted as conv1, conv2_ x, conv3_ x, conv4_ x, and conv5_ x. The convolution kernel size of conv1 is 7 × 7, and the number is 64; the conv2_ x convolution layer has 3 modes with convolution kernel sizes of 1 × 1, 3 × 3 and 1 × 1, and the number of each convolution kernel is 64, 64 and 256 respectively; the conv3_ x convolution layer has 4 modes with convolution kernel sizes of 1 × 1, 3 × 3 and 1 × 1, and the number of each convolution kernel is 128, 128 and 512; the conv4_ x convolution layer has 6 modes with convolution kernel sizes of 1 × 1, 3 × 3 and 1 × 1, and the number of each convolution kernel is 256, 256 and 1024 respectively; the conv5_ x convolutional layer has 3 modulo of convolutional kernel size 1 × 1, 3 × 3, 1 × 1, and the number of each convolutional kernel is 512, 2048 respectively.
Training a network model: inputting the processed image training set and the verification set into the built network model, and continuously optimizing the structural parameters of the model in the training process to finally obtain the trained model weight.
In this embodiment, the training parameters are set to be 100000 times of iteration, batch _ size is 256, the basic learning rate is 0.001, the learning parameters weight _ decay is 0.001, Momentum is 0.8, and step algorithm is adopted, attenuation is performed once every 1000 iterations, Momentum value is 0.8, and one snapshot is output every 4000 iterations.
Step 4, inputting the picture of the hot rolled steel to be detected, shot by the metallographic detection device, of the iron oxide scale of the hot rolled steel to be detected into the trained neural network model by a user, and automatically acquiring the input heat by the neural network modelFe in rolled steel iron scale picture3O4The proportions, distribution regions, and morphological classifications of FeO and eutectoid structures are described. And (3) carrying out image organization recognition by using the trained network: fig. 5 shows an image of a scale structure to be identified according to an embodiment of the present invention, and fig. 6 shows a picture identification result according to an embodiment of the present invention.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions and scope of the present invention as defined in the appended claims.
Claims (3)
1. An intelligent identification method for a hot-rolled steel surface iron scale structure is characterized by comprising the following steps: the method comprises the following steps:
step 1, manufacturing a detection sample of hot rolled steel oxide scale, shooting by using metallographic detection equipment to obtain a hot rolled steel oxide scale picture, and establishing a hot rolled steel oxide scale image sample set after pretreatment;
the manufacturing of the detection sample is to remove stress on the oxide scale of the hot rolled steel through hot inlaying, mechanical polishing and chemical corrosion;
the metallographic detection device is in a backscattering mode of a scanning electron microscope or an electronic probe;
the pretreatment is to adjust the matrix part in the hot rolled steel iron oxide scale picture to be standard white of an RGB color difference space;
step 2, making a structure semantic label graph of each hot-rolled steel iron oxide photo in the marked hot-rolled steel iron oxide image sample set;
step 3, constructing a neural network model for semantic analysis of hot rolled steel oxide scale images, training the neural network model by using the hot rolled steel oxide scale image sample set in the step 1 and the organization semantic label graph in the step 2, and setting the training times for N times to obtain model parameters;
step 4, inputting the picture of the hot rolled steel oxide scale to be detected shot by the metallographic detection device into the trained neural network model by a user, and automatically acquiring Fe in the input picture of the hot rolled steel oxide scale by the neural network model3O4The proportion, distribution area and morphological classification of FeO and eutectoid tissue are described;
step 4.1: reading a jpg image file of the scale of the hot rolled steel to be detected, which is shot by metallographic detection equipment, in a specified folder, and converting the image file into a form of RGB data;
step 4.2: and (3) recognizing a neural network model by using the iron oxide scale structure trained in the step (3), outputting each pixel point of the iron oxide scale structure in the step (4.1) as a feature vector by a last layer of network through a normalized index function softmax, matching a label corresponding to the component with the maximum output vector value, and outputting label file information to obtain a label graph of the image to be recognized, namely the distribution condition of each part of the structure.
2. The intelligent identification method for the scale structure on the surface of the hot-rolled steel product according to claim 1, wherein the step 2 specifically comprises:
step 2.1, Fe of hot rolled steel iron scale picture3O4FeO and eutectoid tissues are subjected to regional division, so that the iron scale phase tissues in the same region are ensured to be the same;
step 2.2, dyeing each area, wherein the embedding material part, the matrix part and the picture shooting information part in the picture are used as background processing and are marked to be black, and the dyed positions are independently output to obtain a tissue semantic label picture;
the dyeing treatment is that the same tissues in the hot rolled steel iron oxide scale pictures on the tissue semantic label pictures of all the sample sets are dyed into the same color;
and 2.3, carrying out image size normalization processing on the structural photo of the iron scale and the semantic tag graph of the corresponding iron scale tissue.
3. The intelligent identification method for the scale structure on the surface of the hot-rolled steel product as claimed in claim 1, wherein the scale structure identification neural network model in step 4.2 is set to 8 layers, wherein the first 5 layers are convolution layers for extracting image features, and the last three layers are anti-convolution layers for image size recovery and logic inference; establishing a pooling layer in the convolution process, and adopting a maximum pooling method of 3-by-3 pooling windows, wherein the step length of the pooling layer is 2, and the pooling layer is used for fusing features and reducing the dimension of the image;
the iron scale structure recognition neural network model adopts ReLU as an activation function, and the expression is as follows:
F(x)＝max(0,x)
wherein x represents a pixel in an RGB channel in the convolution layer, and represents an output value of a neuron in an iron oxide scale structure recognition neural network; and then, inputting the image data into the established network model for training, wherein in the training process, the structural parameters of the network model are continuously optimized, and finally the training parameters are stored as binary files.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010978690.3A CN112085727B (en) | 2020-09-17 | 2020-09-17 | Intelligent identification method for oxide scale structure on surface of hot rolled steel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010978690.3A CN112085727B (en) | 2020-09-17 | 2020-09-17 | Intelligent identification method for oxide scale structure on surface of hot rolled steel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112085727A true CN112085727A (en) | 2020-12-15 |
| CN112085727B CN112085727B (en) | 2023-07-11 |
Family
ID=73737308
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010978690.3A Active CN112085727B (en) | 2020-09-17 | 2020-09-17 | Intelligent identification method for oxide scale structure on surface of hot rolled steel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112085727B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112819814A (en) * | 2021-02-25 | 2021-05-18 | 汇鸿智能科技(辽宁)有限公司 | Metallographic structure identification method and device, computer equipment and storage medium |
| CN112819815A (en) * | 2021-02-25 | 2021-05-18 | 汇鸿智能科技(辽宁)有限公司 | Metallographic structure identification method and device, computer equipment and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101201309A (en) * | 2006-12-14 | 2008-06-18 | 上海梅山钢铁股份有限公司 | Method for measuring hot rolling iron scale and oxide phase surface density thereof as well as usage |
| US20090188589A1 (en) * | 2006-06-01 | 2009-07-30 | Honda Motor Co., Ltd. | High-strength steel sheet and process for producing the same |
| JP2010014721A (en) * | 2009-08-28 | 2010-01-21 | Jfe Steel Corp | Quality control method and manufacturing method of steel plate with oxide film on surface |
| CN101995349A (en) * | 2009-08-10 | 2011-03-30 | 鞍钢股份有限公司 | Corrosion agent of high-steel-level pipeline steel metallographic structure and display method |
| DE102014101318A1 (en) * | 2013-02-05 | 2014-08-07 | Benteler Automobiltechnik Gmbh | Method for producing a Kraftfahrzeugachsbauteils |
| CN105241884A (en) * | 2015-09-29 | 2016-01-13 | 沈阳大学 | Method for measuring ratios of oxide in each phase in hot-rolled carbon steel iron oxide scale |
| CN110363253A (en) * | 2019-07-25 | 2019-10-22 | 安徽工业大学 | A kind of Surfaces of Hot Rolled Strip defect classification method based on convolutional neural networks |
| CN111435121A (en) * | 2018-12-26 | 2020-07-21 | 山西建龙实业有限公司 | Method for detecting and analyzing edge crack defect of hot-rolled steel strip |
-
2020
- 2020-09-17 CN CN202010978690.3A patent/CN112085727B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090188589A1 (en) * | 2006-06-01 | 2009-07-30 | Honda Motor Co., Ltd. | High-strength steel sheet and process for producing the same |
| CN101201309A (en) * | 2006-12-14 | 2008-06-18 | 上海梅山钢铁股份有限公司 | Method for measuring hot rolling iron scale and oxide phase surface density thereof as well as usage |
| CN101995349A (en) * | 2009-08-10 | 2011-03-30 | 鞍钢股份有限公司 | Corrosion agent of high-steel-level pipeline steel metallographic structure and display method |
| JP2010014721A (en) * | 2009-08-28 | 2010-01-21 | Jfe Steel Corp | Quality control method and manufacturing method of steel plate with oxide film on surface |
| DE102014101318A1 (en) * | 2013-02-05 | 2014-08-07 | Benteler Automobiltechnik Gmbh | Method for producing a Kraftfahrzeugachsbauteils |
| CN105241884A (en) * | 2015-09-29 | 2016-01-13 | 沈阳大学 | Method for measuring ratios of oxide in each phase in hot-rolled carbon steel iron oxide scale |
| CN111435121A (en) * | 2018-12-26 | 2020-07-21 | 山西建龙实业有限公司 | Method for detecting and analyzing edge crack defect of hot-rolled steel strip |
| CN110363253A (en) * | 2019-07-25 | 2019-10-22 | 安徽工业大学 | A kind of Surfaces of Hot Rolled Strip defect classification method based on convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| MIKAKO TAKEDA等: "Physical Properties of Iron-Oxide Scales on Si-Containing Steels at High Temperature", 《MATERIALS TRANSACTIONS》, vol. 50, no. 9, pages 2242 - 2246, XP055242339, DOI: 10.2320/matertrans.M2009097 * |
| 赵小龙等: "钢坯在加热过程中的氧化机理及其研究综述", 《轧钢》, vol. 36, no. 6, pages 66 - 68 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112819814A (en) * | 2021-02-25 | 2021-05-18 | 汇鸿智能科技(辽宁)有限公司 | Metallographic structure identification method and device, computer equipment and storage medium |
| CN112819815A (en) * | 2021-02-25 | 2021-05-18 | 汇鸿智能科技(辽宁)有限公司 | Metallographic structure identification method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112085727B (en) | 2023-07-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110110799B (en) | Cell sorting method, cell sorting device, computer equipment and storage medium | |
| CN109509187B (en) | Efficient inspection algorithm for small defects in large-resolution cloth images | |
| CN108009567B (en) | Automatic excrement character distinguishing method combining image color and HOG and SVM | |
| CN111340123A (en) | Image score label prediction method based on deep convolutional neural network | |
| CN111680705B (en) | MB-SSD method and MB-SSD feature extraction network suitable for target detection | |
| CN112085727B (en) | Intelligent identification method for oxide scale structure on surface of hot rolled steel | |
| CN108710893A (en) | A kind of digital image cameras source model sorting technique of feature based fusion | |
| CN112927783B (en) | Image retrieval method and device | |
| CN110895814A (en) | Intelligent segmentation method for aero-engine hole detection image damage based on context coding network | |
| CN115731177A (en) | Method for detecting damage to surface of YOLO v5 steel by introducing attention mechanism and ConvNext network | |
| CN112329771A (en) | Building material sample identification method based on deep learning | |
| CN114399763B (en) | Single-sample and small-sample micro-body paleobiological fossil image identification method and system | |
| Shen et al. | Digital forensics for recoloring via convolutional neural network | |
| CN117541956A (en) | Video semantic feature extraction method based on self-supervised learning | |
| CN112270404A (en) | Detection structure and method for bulge defect of fastener product based on ResNet64 network | |
| CN117095248A (en) | Furnace flame oxygen concentration monitoring method based on generating convolutional neural network | |
| CN113065407B (en) | Financial bill seal erasing method based on attention mechanism and generation countermeasure network | |
| CN115423802A (en) | Automatic classification and segmentation method for squamous epithelial tumor cell picture based on deep learning | |
| CN111754459B (en) | Dyeing fake image detection method based on statistical depth characteristics and electronic device | |
| CN113989567A (en) | Garbage picture classification method and device | |
| CN118052814B (en) | AI technology-based full-automatic specimen pretreatment system and method | |
| CN113112515B (en) | Evaluation method for pattern image segmentation algorithm | |
| Villareal et al. | Patch-based convolutional neural networks for TCGA-BRCA breast cancer classification | |
| CN116071307A (en) | Pavement defect detection model building method, detection method, storage medium and equipment | |
| CN116703848A (en) | Method, readable medium and apparatus for detecting metal surface defect |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |